This report gives the following explanation
"In addition, propensity score matching and inverse probability treatment weighting (IPTW) were used in survival models"
This is using fiddle factors to boost a theory which makes me very suspicious.
The fact that they only found that it was effective in diabetics on statins, but no effect on those not on statins, or who were not diabetic is further weakness in their findings. I have no confidence in this report at all.
Edit to add: A similar study on statins and in-hospitalization mortality in patients with heart failure co-morbidity (ie not diabetes) had the following admission tacked into the small print which shows that this topic is being influenced by interested parties.
Conflict of interest statement
ABD, IB, JR, JJ, SvH, and AVH have no conflicts of interest to disclose; DPM has given talks and attended conferences sponsored by MSD, AstraZeneca and Libytec; SDA reports personal fees from Bayer, Boehringer Ingelheim, Vifor, Servier and Novartis, outside the submitted work; MB has served on the speakers bureau of Abbott/Mylan, Abbott Vascular, Actavis, Akcea, Amgen, Biofarm, KRKA, MSD, Sanofi-Aventis, Servier and Valeant, and has served as a consultant to Abbott Vascular, Akcea, Amgen, Daichii Sankyo, Esperion, Lilly, MSD, Resverlogix, Sanofi-Aventis; Grants from Sanofi and Valean
It seems that the main study I discuss in this post was funded by an NIH government grant and no other funding was declared. B
UT - there was no conflict of interest statement either.